Assessing the Robustness of Web Feature Services Necessary to Satisfy the Requirements of Coastal Management Jennifer Bauer M.S. Geography GISci Certificate June 4, 2012
Presentation Outline!! Background Information!! Coastal Management Issues!! Web-based GIS!! Coastal Web Atlases!! Web Services!! Web Feature Services!! Research Questions!! Methods!! Results!! Discussion and Conclusion
Coastal ecosystems are extremely dynamic and inherently complex ecosystems Photo source: Exxon Valdez Oil Spill Trustee Council GEM (Gulf of Alaska Ecosystem Monitoring and Research) project! There is no one factor, activity, or place for management to consider
Traditional Management Strategies Traditional Ecosystem Management Based Management Strategies Strategies
Building the Toolbox!! Frenzy of tool development by institutions, organizations and agencies to help cope with these problems What are some goals of these tools?!! Incorporate a range of datasets;!! Increase collaboration across stakeholders, jurisdictions, and spatial scales;!! Evaluate management options 1 One popular new tool being used is the coastal web atlas! (Center for Ocean Solutions 2011) 1
What is a Coastal Web Atlas? Photo sources: European Atlas of the Seas A coastal web atlas is a collection of digital maps and datasets with supplementary tables, illustrations and information that systematically illustrate the coast, oftentimes with cartographic and decision support tools, and all of which are accessible via the Internet. 2 (O Dea et al. 2007) 2
Development of coastal web atlases (CWAs) has been driven by imperative coastal management and policy issues!! Photo sources: International Coastal Atlas Network Benefits of CWAs!! Easily accessible;!! Provide access to recent and up to date information;!! Serve as a data catalogue and portal for downloads;!! Incorporation of interactive tools and resources;!! Act as an educational resource;!! Assist with coastal and marine spatial planning 3 (Wright, Dwyer, and Cummins 2011) 3
Augmenting Coastal Web Atlases!! Geospatial Web Services! Modular applications that can be published, located, and invokes across the Web 4! Perform various functions!! Query, search, describe, identify, create, retrieve, etc.! Interface with a variety of applications and services Photo source: Ohio.gov Photo source: Marine Metadata (Vasudevan 2001) 4
Along with providing additional functions, web services also increase interoperability Photo source: akashtrivedi.files.wordpress.com!! Interoperability refers to the ability of heterogeneous systems or system components to communicate, exchange resources, or work together 5, 6! So what exactly does that mean? Allows Coastal Web Atlases to utilize data from different sources and formats for a wide range of functions (Anderson and Moreno-Sanchez 2003) 5 ;(Lassoued et al. 2011) 6
Open Geospatial Consortium (OGC) Created numerous, interoperable, open-source geospatial web service standards 7 Service Interface Standard Input(s) Output(s) Client Functionality Web Map Service (WMS)" Map (vector and/or raster datasets)" Image (GIF, JPEG, etc.)" Request only" Web Feature Service (WFS)" Vector datasets (points, lines, polygons)" XML data (or GML) that includes spatial, metadata, and attribute information" Request, query, and manipulate" Web Coverage Service (WCS)" Raster datasets (pixel-based or a feature bounded in space)" Encoded binary images (GeoTIFF, NetCDF, etc.) and metadata" Request and query" (Reed 2011) 7
Web Services and Coastal Web Atlases: Providing a one-two punch for coastal managers, policy makers, and scientists!! Numerous applications for web services and Coastal Web Atlases!! This study focuses on one approach:! A user s need to explore a specific management concern or question!! Perform complex spatial queries with web feature services (WFSs) to extract relevant information
The core functions of web feature services, as well as Coastal Web Atlases and the Internet, operate based off a client-server architecture The ability to create, author, and access web feature services requires additional components to be added to the basic client-server architecture
Web Feature Service User Requests OGC WFS Standards allow 3 basic operation requests 8 : 1.! Get Capabilities 2.! Describe Feature Type 3.! Get Features http://webserverurladdress/locationofdatastoredaswfs http://webserverurladdress/locationofdatastoredaswfs http://webserverurladdress/locationofdatastoredaswfs http://webserverurladdress/locationofdatastoredaswfs http://webserverurladdress/locationofdatastoredaswfs (Reed 2011) 8
Web Feature Service Limitations Number of integral components required to create, author, and access a web feature service can impact it s robustness Robustness defined: A web feature service s ability to provide accurate and precise data, onthe-fly, in a timely manner and in its entirety for a user, consistently As the role of web feature services in CWAs increases its critical to understand what impacts a WFSs robustness and how that can impact a user
Research Questions Q1. What components are necessary to create a robust web feature services that meets the demands of managers, decision-makers, and scientists to perform critical coastal spatial queries? Q2. How does the robustness of a web feature service affect its performance in executing complex spatial queries?
Wisconsin Coastal Atlas!! Project funded by the University of Wisconsin Sea Grant 9!! Seeks to provide maps, data, and decision support tools to address coastal 9, 10 hazards!! Developed upon the successful framework of the 9, 10, 11 Oregon Coastal Atlas This research is a result of a collaboration between Oregon State University and the University of Wisconsin Sea Grant to evaluate the benefits and limitations of interfacing WFSs with the WCA Photo source: Wisconsin Coastal Atlas (Hart 2011) 9 ; (Ventura et al. 2009) 10 ; (Haddad, Bailey, Wright 2011) 11
Methodology Perform Two Phases of Analysis to Address the Research Questions Phase 1: Research Question 1 Phase 2: Research Question 2 Data sources include Ashland County, Bayfield County, Douglas County, Iron County, Wisconsin Coastal Atlas, and ESRI
Web Feature Service Components Evaluate popular software, hardware, and data components to determine the impact each component has on WFS robustness
Software Components!!Numerous open source and proprietary server and client software applications!!results in a broad range of effects on WFS robustness!!this study will focus on evaluating three popular web mapping servers and desktop GIS applications
Web Mapping Servers MapServer, GeoServer, and ArcGIS for Server 10.0 (henceforth referred to as ArcServer) Python Script Web Mapping Server MapServer County Dataset GeoServer Ashland Bayfield OGC WFS Operation Request (n= 30) ArcServer Douglas Iron Get Capabilities Describe Feature Type Get Feature
Desktop GIS Application Quantum GIS, gvsig, and ArcGIS Desktop GIS Application Quantum GIS gvsig ArcGIS Web Mapping Server MapServer GeoServer ArcServer County Dataset Ashland Bayfield Douglas Iron OGC WFS Operation Request (n= 10) Get Capabilities Describe Feature Type Get Feature
Hardware Components User s Network Speed Capabilities Python Script High Network Speed 1,000 Mbps Low Network Speed 15 Mbps Web Mapping Server MapServer GeoServer ArcServer County Dataset Ashland Bayfield Douglas Iron OGC WFS Operation Request (n= 30) Get Capabilities Describe Feature Type Get Feature
Data Components Impacts of data features, attributes, and metadata!!due to WFSs distributed nature, the amount of data can impact a WFS s reliability and timeliness!!study focuses on three components that can be changed to determine their influence on WFS robustness
Number of Features Representations of real world items Desktop GIS Application Quantum GIS gvsig ArcGIS Web Mapping Server MapServer GeoServer ArcServer County Dataset Ashland 9,192 Features Bayfield 32,843 Features Douglas 47,224 Features OGC WFS Operation Request (n= 10) Get Feature Iron 1,059 Features
Number of Attributes Attributes are data or characteristics of a feature Desktop GIS Application Web Mapping Server Quantum GIS gvsig ArcGIS MapServer GeoServer ArcServer County Dataset Douglas 47,224 Features 33 Attributes Douglas 47,224 Features 6 Attributes OGC WFS Operation Request (n= 10) Get Feature
Metadata Data about data Python Script Web Mapping Server MapServer GeoServer ArcServer County Dataset Douglas No Metadata Douglas Basic Metadata Douglas Mandatory FGDC Metadata Douglas Full FGDC Metadata OGC WFS Operation Request (n= 30) Get Feature
Distributed Spatial Queries Analysis to identify land parcels within 1,000 feet of the Lake Superior shoreline with county parcel data with full features, 6 attributes per feature, and FGDC Metadata Desktop GIS Applications Quantum GIS gvsig ArcGIS
Web Mapping Server Results!! Variations in the creation of each WFS and the WFS XML structure retuned for each request! Impacted file size, download time, overall download speed!! No change in the overall data characteristics of each WFS y=163.34x-4.9844 y=15x-8.5036 y=80.309x-13.971
Desktop GIS Application Results!! Variations in the creation of WFS requests, download time, and overall download speed!! No change in the overall data characteristics of each WFS or downloaded file size
Hardware Results User s Network Speed Capabilities 1400 1200 High Speed Network: MapServer WFS File Size (Mb) 1000 800 600 High Speed Network: GeoServer High Speed Network: ArcServer Low Speed Network: MapServer 400 Low Speed Network: GeoServer 200 0 0 100 200 300 400 500 Average Download Time (s) Low Speed Network: ArcServer
File Size (Mb) File Size (Mb) Data Component Results!! Number of features and attributes mainly impacts a WFS file size and download time!! Accurate return of features and attributes for each test Downloaded <get <get feature> WFS WFS File File Size Size with with Different Different Numbers of Attributes of Features 1400 600 1200 500 1000 400 800 300 600 200 400 100 200 0 Number of Features and Attributes MapServer Iron Ashland GeoServer Bayfield ArcServer Douglas Number of Features Attributes 6 Attributes MapServer GeoServer 33 Attributes ArcServer Average Average Download Download Time Time (s) (s) 90 40 80 35 70 30 60 25 50 20 40 15 30 10 20 10 5 0 Average Average Download Download Time Time of of <get <get feature> feature> WFS request WFS request with Different with Different Numbers Numbers of Attributes of Features MapServer Iron Ashland GeoServer Bayfield ArcServer Douglas MapServer 6 Attributes GeoServer 33 Attributes ArcServer
Data Component Results Metadata!! No effect on the number of features or attributes downloaded!! No or small variation in file size and download time between all levels of metadata File Size (Mb) 1400 1200 1000 800 600 400 200 0 Downloaded <get feature> WFS File Size with Different Levels of Metadata No Meta Basic Meta FGDC Mand. Full FGDC MapServer GeoServer ArcServer Average Download Time (s) 90 80 70 60 50 40 30 20 10 0 Average Download Time of <get feature> WFS request with Different Levels of Metadata No Meta Basic Meta FGDC Mand. Full FGDC MapServer GeoServer ArcServer
Distributed Spatial Queries Result!! No effect on the results of the spatial query!! Only impacts were on the user: such as tool location and tool parameters County Number of selected features Total Value (dollars) Ashland 1,787 158,419,410 Bayfield 2,441 108,622,950 Douglas 1,671 39,260,500 Iron 91 780,500
Implications of Results!! Study suggests that all components had an impact on WFS robustness! Predominantly impacted by server software and network speed! Greatest impacts were on WFS file size and download time!! Selecting WFS components should be based off a potential users needs!! Robustness of a WFS did not affect the accuracy of a distributed spatial query Photo source: Jim Bauer
Conclusion!! Web Feature Services:! Reliable, on-the-fly spatial coastal datasets! Flexible! Utilized in a variety of coastal management applications! Increase the use of a Coastal Web Atlas Photo source: Wisconsin Coastal Management Program
Acknowledgements A special thanks to my advisor, Dr. Dawn Wright, and my graduate committee, Dr. Julia Jones and Dr. David Hart. Thanks to Dr. Jim Graham for his help with web services and python scripting. Huge thanks for the support from my follow Rogues in Davey Jones Locker and well as my OSU cohort, and friends scattered across the U.S. Thank you to my family for endless support and encouragement. And a most grateful thanks to my husband, David, for his endless support and help during this entire process
Questions? Photo source: Wisconsin Coastal Management Program "#5#"-&#$!! Anderson, G., and R. Moreno-Sanchez. 2003. Building Web-Based Spatial Information Solutions around Open Specifications and Open Source Software. Transactions in GIS 7: 447-66. ArcGIS Desktop 10.0. 2011. ESRI (Environmental Systems Research Institute), Inc., Redlands, CA. ArcGIS for Server 10.0. 2010. ESRI Inc., Redlands, CA. &67869!:;9!,<6=7>!$;?@A;7>B!CDEFFGB!/6<H>H;7!I@HJ6K!$6?6<A7I!J6<H>H;7!>@LL;98!8;;?>!:;9!M=9H76!>L=A=?!L?=77H7IB!*N6!O;;J>!+7>A8@86!:;9!8N6!#7PH9;7M678Q!$8=7:;9J!)7HP69>H8RQ!&=?H:;97H=B!%P=H?=S?6!=8K!NTLKUUVVVB<67869:;9;<6=7>;?@A;7>B;9IU L@S?H<=A;7>WSRW8;LH<UFXE GeoServer 2.1.3. 2011. GeoServer and the Open Source Geospatial Foundation, Beaverton, OR. gvsig 1.11.0. 2011. gvsig Association and the Open Source Geospatial Foundation. Valencia, Spain. Haddad, T., Bailey, R. J., and D. J. Wright. 2011. Oregon, USA. In Coastal informatics: web atlas design and implementation, ed. D. J. Wright, N. Dwyer, and V. Cummins, 91-104. Hershey, PA: IGI Global. Hart, D. 2011. Wisconsin, USA. In Coastal informatics: web atlas design and implementation, ed. D. J. Wright, N. Dwyer, and V. Cummins, 145-55. Hershey, PA: IGI Global. Lassoued, Y., Pham, T. T., Bermudez, L., Stocks, K., O'Grady, E., Isenor, A., and P. Alexander. 2010. Coastal atlas interoperability. In Coastal informatics: web atlas design and implementation, ed. D. J. Wright, N. Dwyer, and V. Cummins, 53-78. Hershey, PA: IGI Global. MapServer 5.4.2. 2009. MapServer Project Steering Committee and the Open Source Geospatial Foundation, Beaverton, OR. O Dea, L., Cummins, V., Wright, D., Dwyer, N. and I. Ameztoy. 2007. Report on Coastal Mapping and Informatics Trans-Atlantic Workshop 1: Potentials and Limitations of Coastal Web Atlases. University College Cork, Ireland. http:// workshop1.science.oregonstate.edu/final_rpt (last accessed 07 May 2012). Quantum GIS 1.7.4. 2012. Quantum GIS Project Steering Committee and the Open Source Geospatial Foundation, Beaverton, OR.. Reed, C. N. 2011. The open geospatial consortium and web service standards. In Geospatial Web Services: Advances in information interoperability, ed. P. Zhao and L. Di, 1-16. Hershey, PA: Information Science Reference. Vasudevan, V. 2001. A web service primer. http://webservices.xml.com/pub/a/ws/2001/04/04/webservices/index.html (last accessed 05 May 2012). Ventura, S., Hart, D., Wiegand, N., Buckingham, T., Stoltenberg, J., Wortley, A., Wright, D., and T. Haddad. 2009. The Wisconsin coastal atlas: Building a coastal spatial data infrastructure for Wisconsin [a proposal to the University of Wisconsin Sea Grant Institute]. 34. Wisconsin. Wright, D. J., Dwyer, N., and V. Cummins. 2011. Coastal informatics: Web atlas design and implementation. Hershey, PA: Information Science Reference.